mircea_popescu: all this because they really wanted to be as cool as the euros and didn't comprehend that faking it till you make it just gives you saturnism
mircea_popescu: then it would be.
mircea_popescu: except if everyone had plutonium induced braindamage from the fashion of eating plutonium pills
asciilifeform: well not as such. but he very much saw it as something ~not~ created by teenagers in mother's basement
mircea_popescu: even if he may or may not believe any particular building is the result of some people who did some calculations in private for money.
asciilifeform: naggum was an old-worlder. in the very specific sense of taleb et al
assbot: Logged on 20-02-2016 22:20:40; mircea_popescu: moe finds bitcoin in 2056, long after the republic prevailed, through his gf at the time. he steals her script, copies her .wot directory and now has a node!
mircea_popescu: "(4) you could communicate your desire for a system-wide improvement to someone who could put it in the official sources and you could add your changes conditionally." << see specifically how he manges the very pointed difference discussed in http://log.bitcoin-assets.com/?date=20-02-2016#1411186
asciilifeform: these are not actually separate, causally
mircea_popescu: that the concept is poorly implemented is a separable and in fact separate problem from that the concept is broken
asciilifeform: the promises made by a c machine are false promises.
mircea_popescu: asciilifeform the sufficient and necessary concrete is readily available and of an abstract kind. i needn't any actual calcium carbonate to say various things about various bits of code.
asciilifeform: otherwise you might cat /dev/random > /dev/mem
mircea_popescu: currently our substance is drawn specifically from the fact that for the past 2-3 decades, these twerps approached the problem of functionality ("the software doesn't do enough") with nary a clue about the problem of feedback ("the software does too much")
asciilifeform: this 'never do' exists only if you burn into rom and then pour concrete around it.
mircea_popescu: how so. how the fuck are you EVER going to sign off on any code when it is not possible to say what function call will never do ?
mircea_popescu: what he's talking about it is fractal they shall be delivered thing
asciilifeform: then again he did pay for src licenses
mircea_popescu: the principal specification of any machine, not JUST computers, not just software stacks, not just lines of code, is to specify what it can never do.
mircea_popescu: i do not WISH to run an "open" which may acquire unspecified powers in the future if the caller rubs its belly just right.
mircea_popescu: nevertheless, the "solution" he contemplates is of the level of "because inept us-style corporatism displays all the sechelae you'd expect if you got a bunch of farmers to say how cars should look, let's everyone run in traffic!"
mircea_popescu: he is correct in that foss is mostly a problem unto itself, making the previous situation worse rather than better.
mircea_popescu: asciilifeform for the record this " the reason that source access is needed is the same as the cause of the problem: the inability to modify system behavior non-intrusively. if one could modify OPEN on any given system to acquire new behavior _without_ compiling a new OPEN or, even if you did, without re-linking the entire system (thus the need for dynamic linking and similar solutions, too), source code access would
gernika: phf: Had a terrible vegetarian thanksgiving dinner at Breitenbush. I did see a naked woman there though, as a teenager.
assbot: Jared Carrabis on Twitter: "Welp. Here it is. Via the @BostonGlobe, the picture that everyone's going to be talking about tomorrow. https://t.co/FbWPrm8DfH" ... ( http://bit.ly/1L49wg0 )
phf: i used to do the whole 30 minute run through winter forest for a cold plunge in прорубь thing growing up, so i like hotsprings in winter. you get a pretty reasonable crowd around this time too, no idea how it is in summer
mircea_popescu is not terribly fond of hot springs, for some reason they attract old folks.
phf: in fact i'm going to breitenbush two weeks from now, in the ben_vulpes area
phf: fwiw cascade range has a bunch of fantastic hot spring places. there's one next to Seattle, one next to portland, i mean of the the ones i've been too. but there's also numerous holes of course
phf: oh, is that some other bathhouse in nyc?
polarbeard: my daily work involves the rightfully hated javascript
mod6: Ok Gentlemen, I've gotta run for a few here. Jacmet: Thanks again for stopping in. When you're ready with your new key, stop on back and we'll help you get setup in the WoT. Cheers!
mircea_popescu: moreover, it's marked for replacement. just, iffy thing to replace, the crypto lib.
Jacmet: mod6: yes, the download infrastructure in buildroot will only download if the tarball isn't already available in the dl dir, or you can point it to a local mirror
mod6: <+Jacmet> mod6: and why the manual download of openssl/db/boost instead of just letting Buildroot do it? << actually, one thing I think we're considering is doing a modification to buildroot so that we don't pull its deps over the wan via rsync. is there a way, currently, to just place all of its deps somewhere so it won't rsync?
Jacmet: asciilifeform: E.G. updated openssl, there's been some security fixes since 1.0.1g
mircea_popescu: Jacmet tmsr doth not believe in upgrades. on the contrary, the versions here count towards 0.
mod6: np. well, we've been working for quite a while now to get all of this tooling setup the way we need it. we haven't had a reason to move forward with the version.
Jacmet: mod6: and why the manual download of openssl/db/boost instead of just letting Buildroot do it?
mircea_popescu: mod6 got a link to the current sh ?
phf: actually i don't know how to call them..
phf: while early in the game, stable, experimental, uploads and deprecated
gribble: The operation succeeded.
assbot: Logged on 02-12-2015 04:13:47; BingoBoingo: "On a personal "hate ramen noodles and tuna" level, I agree. But my good-for-project-good-for-the-world side says the OpenBSD Foundation is more effective at growing the contribution pie and in particular funding the hackathons where great work happens." - Theo De Raadt
mod6: <+mircea_popescu> make a 4kb key and use it rather than the strange md5 lists or what's that. << RSA too.
assbot: Logged on 16-02-2016 16:36:23; asciilifeform: now the other thing about openbsd is the incident with theo and that microshit flunkie
asciilifeform: Jacmet: if you open a nonprofit org, esp. in usa, you will be surrounding yourself with, and eventually giving control over various aspects of the project, to people who do not work in good faith.
ben_vulpes: whats the connection between key and file hash algos?
ben_vulpes: Jacmet: another thing worth pointing out are that sha1 and md5 are obsolete
asciilifeform: Jacmet: i dare say that linux is only meaningfully alive today in the form of buildroot.
asciilifeform: mircea_popescu: can't seem to find it in the logz
mircea_popescu: x.org just recently lost their only asset because of inept "non profit" shenanigans.
mircea_popescu: so don't make the mistake.
mircea_popescu: pick a strong leader in the form of a person, build your thing atop pgp sigs and forget the nonsense fiat.
mircea_popescu: that's a horrible idea if the wya such things work out in practice is any guide.
asciilifeform: Jacmet: the background: i originally picked up buildroot to do an arm system build for an obscure box. and then discovered that it is also the only practical means of compiling with musl instead of glibc, for any system
Jacmet: asciilifeform: ok, I would be *VERY* interested in comparing them
Jacmet: asciilifeform: what was the difference between the bad/good tarball? Real rogue code or just interrupted download?
mircea_popescu: asciilifeform where was the line/dpaste ? dja recall ?
Jacmet: asciilifeform: really? We never had reports on the ml or on irc
Jacmet: mircea_popescu: true. I'll get the new key cross signed by the other devs at our next dev days and start using it
mircea_popescu: that would be the one problem - the thing is a very valuable security asset, should be better protected.
Jacmet: and yes, the current gpg key used for buildroot signing is getting quite old
ben_vulpes: Jacmet: buildroot is working rather niftily.
mircea_popescu: and buildroot with musl prevented us from bathing in the same humiliation EVERYONE else ended up in with the recent gethostbyaddr thing
mircea_popescu: we are : it's become the replacement harness for bitcoin.
asciilifeform: 'if all you have is Unix, certain problems are so staggeringly tall that it takes an extra special genius to see what the _real_ problem is because all the people who are trying to climb it make it look as if doing something else is stupid.' << naggum delivers.
assbot: Logged on 21-02-2016 16:54:08; asciilifeform: we have finessed the halting problem !1111
ben_vulpes: i read 'tc' in the sense of 'can express arbitrary actions', in the vein of the old joke about programming in $shitlang: "hey, it's turing complete, we can make it do whatever we need"
mircea_popescu: "arbitrary text transformations" but only between these two sets we carefully curated.
asciilifeform: the language is gargantuan, just enough parser to ~compile~ is likely to take you a decade to write.
jurov: then it couild also parse c++
asciilifeform: sure they do.
asciilifeform: betcha somebody's very busy stuffing the winblowz crud back in, for instance.
jurov: so that descendants will list code snippets the actually depend on
asciilifeform: if only it were so simple. do you notice what most programming types do when they run into a forest ?
mircea_popescu: he's bound to run into a shitton of resistance of the medium in short order with that thing.
asciilifeform: it needs to come out of the 'resistance of the medium' from the sort of thing we find ourselves doing with regularity.
mircea_popescu: which is what the first pass of this convo actually exactly was, iirc.
mircea_popescu: ok, you know, say that! "i want a homoiconic thing with lambda", rather than "tc"
jurov: lol first me tormenting alf with ed, now they awk
mircea_popescu: asciilifeform what other than regexp do you hate in awk ? and being gnucpp, of course.\
asciilifeform: enough with the arbitrary nonorthogonal craplangs.
jurov: no the ide was patchmachine takes *filesystem* and ditto
phf: i think the idea is to have a slightly more constrained awk. patchmachine takes string, manipulates it, returns new string
mircea_popescu: good luck getting the fps to do so, however.
asciilifeform: i suppose, like every other time before, i am doomed to demonstrate ~by example~...
mircea_popescu: in that it's "turing complete in a purely theoretical sense and with none of the practical overhead a practical implementation would imply except in those cases where it is clearly and truly needed".
asciilifeform: all i ask for is that vtron be able to specify arbitrary transformations, e.g., 'replace every letter a with b', rather than the +++/--- imbecility.
mircea_popescu: but a very... haskellite, even, theoretical notion.
mircea_popescu: and in the world's most amusing porn news, "pregnant male doctor grabs boob" http://41.media.tumblr.com/cbf106747cae91ec1e7fa6bbad5fcbe9/tumblr_mlvjrxMRnl1qlne6uo1_1280.jpg
asciilifeform: these, i think, are meant to be debug hooks
asciilifeform: there is a demo, which was supplied with the original patch, that gets block count, and another knob that quits.
asciilifeform: there are no useful hooks yet.
asciilifeform: ben_vulpes: much more generally, and much later - as potential replacement for entire chunks of the rotten flesh.
asciilifeform: ben_vulpes: multifold point. most realistically, and in the short term, to abolish the json rpc crapolade.
mircea_popescu: tied some ropes to it and pulling on them, technically yes "this pulls the earth the wrong way". practically, tis alpinism.
mircea_popescu: there's nothing wrong with treating your baby as a chunk of bedrock, you know.
asciilifeform: and i am all for the fella figuring things out, so long as he understands this, and the ~why~.
asciilifeform: well it pull in the direction precisely opposite to shiva.
mircea_popescu: won't be accepted in trb by any means, but i dun see anything wrong wiht the process.
mircea_popescu: so the guy is more familiar with cpp, naturally wants to isolate the weird he's taking in and take it in at controlled rate.
asciilifeform: don't need to. but see the thing in question, imho it is more or less the perfect example of what not to do.
mircea_popescu: god help the man that takes a nap during republican times.
mircea_popescu: i shall now proceed to read the scrollback!
asciilifeform: phf: only the first time. it then turns into a steel beam you can use to build proper towers.
asciilifeform: ^ apparently the whole thing. 1 page.
mod6: haha. also gives you the shits.
ben_vulpes: but the phood is phr33!
gribble: The operation succeeded.
mod6: ;;later tell pete_dushenski here's that change log you wanted: http://therealbitcoin.org/ml/btc-dev/2016-February/000213.html
asciilifeform: or where, found on the street ?
ben_vulpes: to be frank, i have no idea how this gentoo image was even built in the first place.
asciilifeform: pretty much everything we have, we have because someone ran the motherfucking gnomes to ground, each time.
asciilifeform: and other evils
asciilifeform: it is how we discovered the libnss thing
phf: ben_vulpes: i'm building it from inside Clion, so i'm not sure how easy it is to do cmake otherwise, i can share the diff though
ben_vulpes: mod6: under gentoo the vm simply gives up on life
ben_vulpes: here's the last line, asciilifeform: x86_64-therealbitcoin-linux-musl-gcc: internal compiler error: Segmentation fault (program cc1)
ben_vulpes: and in other nyooz of the weird and strange saga of ben_vulpes endeavoring to set up a reference implementation on his heathen macintosh, buildroot fails to compile in linux virtual machines of all sorts of flavors
mod6: no super rush there.
mod6: so once we're happy with this build-script, etc. we cna update that wiki, then we're back in sync. then I can take some time to get the makefiles stuff all organized.
mod6: 9 connections, there soemthing i don't see everyday.
shinohai: Me too did u get my pm with the address to node stats?
mod6: im making a `typescript' of the output as well. will post later when its complete.
assbot: Logged on 21-02-2016 20:08:29; phf: well, init.scm is read on each connection, since there's no persistence
asciilifeform: phf: http://log.bitcoin-assets.com/?date=21-02-2016#1412300 << as stated in the comments, i regard this as a serious loose end that must be tied before anyone can even contemplate using a shivatronic trb in the battlefield
mod6: the link is above if anyone wants to give it a try.
mod6: ah, ok good deal shinohai. I updated the script so that it doens't even pull that .sign file. we'll just rely on the sha512 that I hvae, alf has, and ben has.
mod6: then that makes a bit more sense.
jurov: most perl scripts meddle with $SIG so they worked fine
jurov: but perl failed to execue them!
jurov: THEY DID WORK
mod6: my mind is blown. these are C binaries.
jurov: yay just recompiled perl without thread support, that did the trick
mod6: asciilifeform or trinque have any thoughts for the man?
mod6: because then your pwd would be /
jurov: mod6 well, on another box i updates glibc to same version and perl works fine
phf: well, init.scm is read on each connection, since there's no persistence
asciilifeform: phf: possibly it ought to die before the db is opened.
asciilifeform: jurov: look in the portage cache, possibly. but you might have to diff with previous.
phf: (i.e. no reason to corrupt db in the process)
mod6: it'd be like doing the same thing in a bash script: PWD=`pwd`; echo $PWD;
asciilifeform: phf: and this is the correct behaviour (IF SHIVA IS ENABLED), just needs to happen cleanly and humanely. does it ?
mod6: pwd is used iirc in v.pl, but only in the contex of it being a shell out - your environment should execute this and return the result to a scalar.
jurov: i updated glibc with the patch
asciilifeform: or did you... update the box.
mod6: take a look, i've removed 3 lines: 1. in the comment section where it lists PK's key fingerprint at the top 2. the curl that pulls the sign file for buildroot-2015.05.tar.gz.sign 3. the gpg command that verifies buildroot-2015.05.tar.gz.sign
mod6: so as far as the build-bitcoind-V99995.sh is concerned. let's just check that hash and call it a day then.
mod6: anyway, an open line of comms to these folks could help us -- especially going forward as we basically might need to "roll our own" so that it doesn't pull the deps via rsync.
mod6: SHA1 & MD5 are broken are they not?
mod6: just thinking that hey, cant hurt anything to enlighten the guy that these hashes that he's signing, even up to this current month are obsolete.
ben_vulpes: asciilifeform produced legacy hashes, they match the ones i have on disk and the ones you have right?
ben_vulpes: "the army we have"
mod6: or at least, just the one we're using.
mod6: what we really aught to do is write to that guy, get him to join the wot with a 4096 bit RSA key, and have him resign all of the bundles.
asciilifeform: and their cached crapolade.
asciilifeform: and their tars.
mod6: asciilifeform:this is the one that we're pulling
asciilifeform: in other nyooz, from the depths, http://www.newsbtc.com/2016/02/21/xapo-announces-plans-to-upgrade-to-bitcoin-classic
ben_vulpes imagines stan's face shooped onto the manul
mod6: im starting to think that our SHA512 sum of this artifact is better than whatever they've got.
mod6: why even bother?
mod6: heh, these guys built this huge thing, but their sign files only have SHA1 & MD5, even the most recent 2016.02
mod6: oh nm, looks like those are only there for an RC version?
mod6: guess that only started with the version /after/ our version 2015.05
mod6: in here there seem to be .sign files for the .tar.gz's, but they seem to have .gpg's for the tar.bz2's
ben_vulpes: ha wow did the shitgnomes snip that mod6 ?
phf: yeah, korean places are not bad, but i know only of one that caters specifically to koreans (i.e. good heat), a lot of them are western oriented so massage and facials and such
kakobrekla: im fine with shower only as long as you can set the temp to whatever you want
phf: there's literally 4 places in the entire virgnia-connecticut stretch. maybe there's something up in Massachusetts. there's "east european" places ran by russians but with a lot of turks, etc. and then there's korean baths. but koreans frown on cold plunges, so they would usually have a tepid shower/pool
mod6: ben_vulpes: yeah, i dunno what the deal is with it, seems to not work exactly as expected.
kakobrekla: really no saunas? whats the deal with that?
phf: as soon as i can move again, i'm going to make it to russian&turkish baths in nyc. they don't really do saunas on the east coast, but that baths place is awesome
kakobrekla: meanwhile i was the only person in sauna yesterday evening
phf: kakobrekla: i went camping in the Appalachian last weekend, -20c w/ winds, but it was awesome. i did just enough trash talking about russian blood to come down with a flu by thursday...
mod6: anyway, i guess i don't mind, i can add a part where we check the SHA1 & MD5 of the buildroot artifact and then continue to verify that .sign file.
mod6: danielpbarron: ok thanks for the update
mod6: i think this was actually discussed quite some time ago. i dont remember what the sentiment was exactly. but i think it was like "meh"
kakobrekla had a blast snowboarding the prime slopes of Dolomites region of Alps
phf: ben_vulpes: they are like md5 and sha1
ben_vulpes: mod6: perhaps hard code the hashes from korsgaard's signature and verify those?
mod6: i guess, the other good news to that is you wont need to import that rando's guy's key into your keyring. still would need mine, alf & trinque's tho.
mod6: so with that build script, i'm thinking just take out the curl & --verify of that stupid .sign file.
phf: added patch highlight to btcbase (got tired of trying to visually locate patches in the graph)
asciilifeform: gotta do a few things to the mains wiring, i (and zoolag) will be down for 30 min. to an hour.
asciilifeform: the fact that reiser was jailed cemented it as permanently high-quality, no one wants to touch it, no 'progressors' progressate at it, no gnomes, no popularizers!111, etc.
asciilifeform: and much as i loathe the idea, we will probably have to nail down our own linux distro.
asciilifeform: instruct new users that they must operate on a sane fs.
asciilifeform: anyway this particular crapocalypse is avoidable, at least on our planet. BAN any heathen fs.
jurov: lol looky now i'm the one to predict crapocalypse
asciilifeform: see also the gcc5 threads
asciilifeform: gotta love these 'optimizations'
asciilifeform: srsly, we are there already ??
assbot: Logged on 21-02-2016 16:07:32; mircea_popescu: there's been this weirdo trend of (mostly anglo) girls getting worried about their labia minora being "wrong", mostly on account of thgere being a LOT of natural variation on that score, and anglo chicks being all weird and insecure and sexually inexperienced. plus of course the crazed marketing society, that feeds into all of the above.
asciilifeform: any future folks who might think they are being clever with, e.g., nonterminating vtron program.
jurov: also, you imply basically to expose all filesystem primitives to scheme, are they comprehensively and soundly defined?
asciilifeform: we have finessed the halting problem !1111
asciilifeform: there is exceed-bound-and-i-negrate
asciilifeform: jurov: the magic trick is that there is no such thing as unbounded runtime on my box.
jurov: asciilifeform: either it's 'can do all but X' or turing complete and thus with potentially unbounded runtime
asciilifeform: it is almost exemplary of the opposite of what we need here.
jurov: http://log.bitcoin-assets.com/?date=21-02-2016#1411948 << have you looked into 'ed' diffs? but yes, they can't rename directory either.
polarbeard: it works well, but still simple hooks because only a single callback may be registered per hook, not sure yet if there will be need for callback management
polarbeard: it actually stores the callbacks
asciilifeform: but i can't promise this, and can't be bothered to test, for reasons already described.
asciilifeform: from my naked eye reading, your thing will run until the first gc cycle.
asciilifeform: the magic of 'v' is that polarbeard does NOT have to agree with me!111
asciilifeform: but yes, in the long run it is intended as a prosthesis to take over where the brain damage of the original is incurable.
asciilifeform: nobody asked you to. but try also refrain from translating shiva to cpp, and spreading the leak metastases into it ?
asciilifeform: 1) to have the cpp content of trb asymptotically go to ZERO with time
asciilifeform: (not to mention, bloating the code, but this is secondary)
asciilifeform: is that by 'marshalling' the pointers, you are introducing trb's leakages into tinyscheme.
asciilifeform: see, the big thing here,
asciilifeform: these i will happily read, and sign.
asciilifeform: anything that is not going there, must answer why.
asciilifeform: for another, bug fixes are their own patches, one per bug, that must contain nothing else.
polarbeard: so, what you're going to do with ts then?
polarbeard: I'm just encapsulating the functionality